From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 48) id B2467386F000; Mon, 21 Dec 2020 15:08:58 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B2467386F000 From: "jiawei at iscas dot ac.cn" To: gcc-bugs@gcc.gnu.org Subject: [Bug target/97417] RISC-V Unnecessary andi instruction when loading volatile bool Date: Mon, 21 Dec 2020 15:08:58 +0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: changed X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: gcc X-Bugzilla-Component: target X-Bugzilla-Version: 10.2.0 X-Bugzilla-Keywords: missed-optimization X-Bugzilla-Severity: normal X-Bugzilla-Who: jiawei at iscas dot ac.cn X-Bugzilla-Status: NEW X-Bugzilla-Resolution: X-Bugzilla-Priority: P3 X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org X-Bugzilla-Target-Milestone: --- X-Bugzilla-Flags: X-Bugzilla-Changed-Fields: Message-ID: In-Reply-To: References: Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/ Auto-Submitted: auto-generated MIME-Version: 1.0 X-BeenThere: gcc-bugs@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-bugs mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 21 Dec 2020 15:08:58 -0000 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D97417 --- Comment #54 from jiawei --- Hi Jim. I had finished the test on the benchmark Coremark-pro.And it shows that the patch doesn't accidentally increase code size. This test with the args "XCMD=3D'-c4' certify-all", and the result shows fo= llow: WORKLOAD RESULTS TABLE(origin gcc-10.2.0 from upstream compiled with -mabi= =3Dlp64 -march=3Drv64imafdc) MultiCore SingleCore=20=20= =20=20=20=20=20=20=20=20=20 Workload Name (iter/s) (iter/s)=20=20= =20 Scaling ----------------------------------------------- ---------- ---------- ---------- cjpeg-rose7-preset 119.05 46.08=20=20= =20=20=20=20 2.58 core 1.10 0.28=20=20= =20=20=20=20 3.93 linear_alg-mid-100x100-sp 35.19 8.92=20=20= =20=20=20=20 3.95 loops-all-mid-10k-sp 1.77 0.46=20=20= =20=20=20=20 3.85 nnet_test 1.69 0.51=20=20= =20=20=20=20 3.31 parser-125k 43.01 15.62=20=20= =20=20=20=20 2.75 radix2-big-64k 257.93 56.84=20=20= =20=20=20=20 4.54 sha-test 156.25 57.80=20=20= =20=20=20=20 2.70 zip-test 81.63 26.32=20=20= =20=20=20=20 3.10 MARK RESULTS TABLE Mark Name MultiCore SingleCore=20=20= =20 Scaling ----------------------------------------------- ---------- ---------- ---------- CoreMark-PRO 2669.64 796.33=20=20= =20=20=20=20 3.35 WORKLOAD RESULTS TABLE(Add patch one -- Auto-extend Patch) MultiCore SingleCore=20=20= =20=20=20=20=20=20=20=20=20 Workload Name (iter/s) (iter/s)=20=20= =20 Scaling ----------------------------------------------- ---------- ---------- ---------- cjpeg-rose7-preset 131.58 45.87=20=20= =20=20=20=20 2.87 core 1.08 0.28=20=20= =20=20=20=20 3.86 linear_alg-mid-100x100-sp 35.49 8.75=20=20= =20=20=20=20 4.06 loops-all-mid-10k-sp 1.75 0.46=20=20= =20=20=20=20 3.80 nnet_test 1.68 0.51=20=20= =20=20=20=20 3.29 parser-125k 54.05 15.62=20=20= =20=20=20=20 3.46 radix2-big-64k 257.80 65.57=20=20= =20=20=20=20 3.93 sha-test 153.85 57.47=20=20= =20=20=20=20 2.68 zip-test 76.92 26.32=20=20= =20=20=20=20 2.92 MARK RESULTS TABLE Mark Name MultiCore SingleCore=20=20= =20 Scaling ----------------------------------------------- ---------- ---------- ---------- CoreMark-PRO 2737.52 806.42=20=20= =20=20=20=20 3.39 WORKLOAD RESULTS TABLE(Add all patches -- Auto-extend Patch & untested fix = to use instead of levy's combine.c patch) MultiCore SingleCore=20=20= =20=20=20=20=20=20=20=20=20 Workload Name (iter/s) (iter/s)=20=20= =20 Scaling ----------------------------------------------- ---------- ---------- ---------- cjpeg-rose7-preset 120.48 45.66=20=20= =20=20=20=20 2.64 core 1.08 0.28=20=20= =20=20=20=20 3.86 linear_alg-mid-100x100-sp 35.61 8.77=20=20= =20=20=20=20 4.06 loops-all-mid-10k-sp 1.76 0.46=20=20= =20=20=20=20 3.83 nnet_test 1.68 0.51=20=20= =20=20=20=20 3.29 parser-125k 46.51 15.62=20=20= =20=20=20=20 2.98 radix2-big-64k 257.33 65.18=20=20= =20=20=20=20 3.95 sha-test 153.85 57.47=20=20= =20=20=20=20 2.68 zip-test 83.33 26.32=20=20= =20=20=20=20 3.17 MARK RESULTS TABLE Mark Name MultiCore SingleCore=20=20= =20 Scaling ----------------------------------------------- ---------- ---------- ---------- CoreMark-PRO 2691.95 805.68=20=20= =20=20=20=20 3.34 The csibe test is still modifying,and will test after it works on riscv-gnu-toolchain.=